In the diagnosis of pulmonary embolism some centres using 133Xe for comparison with multiple view 99Tcm perfusion images perform only single-breath posterior view ventilation scans. The purpose of this study was to test the reliability of the posterior view ventilation scan in the detection of lobar and segmental defects in ventilation. Occluding balloon catheters were placed in lobar and segmental bronchi during fibreoptic bronchoscopy to produce defects of known anatomical location and size in normal volunteers. Subjects breathed 81Krm/air during the occlusions and images were acquired in the posterior, posterior/oblique and lateral projections. The posterior view images were classified by three experienced nuclear medicine physicians as normal or abnormal. If abnormal, the observers were asked to state which lobe or segment was involved. Segmental defects were missed in 28% of scan readings. Segmental defects were detected but incorrectly sited in 50% of readings and correctly sited in only 22% of readings. The posterior view scan with a defect involving the entire lingula was judged to be normal by all observers. Defects involving the right and left lower lobes were underestimated. We conclude that ventilation scanning techniques that assess the distribution of ventilation in the posterior view alone are unreliable in the detection of segmental and lobar defects, and are likely to increase the false positive rate in the diagnosis of pulmonary embolism.